Functional MRI (fMRI) has had perhaps the strongest influence of any neuroscience method on the public conception of how the brain operates. For the past decade or longer, colorful maps have adorned science articles in the popular press, raising the public awareness of neuroscientific progress. The fact that it is possible to look inside the head of an awake, behaving person to discover not only the detailed structure of their brain but also, to some extent, what they are thinking, is truly in the sphere of science fiction. At the same time, there is a limit to how much useful information conveyed with an activity map, and the trajectory of the fMRI field is presently unclear. One of the major limiting factors of this method is the fact that its substrate for estimating brain activity is in a hemodynamic blood flow-based signal, which is quite a bit removed from the electrical activity of neurons that are at the core of our cognition. While the measured fMRI signal at a given point in the brain is known to reflect local neural activity, and is in this sense better than other human methods such as EEG and MEG, it does so in a manner that is very coarse, both in its timing and in its spatial precision. There is a general need to understand how measured fMRI signals might relate to underlying neural activity. Understanding this fMRI/neural relationship can take multiple forms. In some laboratories, experiments are aimed at carefully determining what types of neurons and molecules serve as the actual mediators for regulating blood flow. Such experiments ask, what are the biological principles, and specific mechanisms, that determine neurovascular coupling? Our approach is somewhat different and focuses on more targeted and practical questions. One such question asks how does the firing of a single neuron, or a population of neurons, contribute to the measured fMRI response within a small volume of the brain, or voxel? Consider the following: When measuring brain activity with fMRI, the smallest unit of brain activity--the electron so to speak--is a voxel that is at least a millimeter in each dimension, often much more. As neurons are much smaller and packed in high density, each voxel in the brain contain hundreds of thousands of individual neurons. We know from direct recordings of such neurons that neural responses in such a volume are diverse. One can see this local diversity when observing spontaneous brain activity, or even under conditions of visual stimulation. For example, neighboring neurons within a single voxel respond very differently when a subject is viewing a video sequence. Yet, if all of the neurons respond differently at a given moment in time, then how can the hemodynamic signal from the point in the brain be interpreted in terms of neural firing? Would the fMRI signal reflect the mean neural activity in the population? Are there particular, special neurons that determine the fMRI signal? And would the relationship remain fixed, both during periods of stimulation and rest in which visual input is minimized? Over the past year, we have published four papers related to the investigation of fMRI signals, their relationship to neural activity, and tools to map activity effectively onto specific brain areas. In one paper (Chang et al (2016) PNAS), we investigated the manner in which brain arousal was simultaneously reflected in behavioral, electrophysiological, and fMRI measures. Specifically, we found that behavioral changes in arousal, matched by corresponding changes in the cortical field potentials, were yoked to hemodynamic fluctuations in an identified multicomponent network, or template, that could be used to read out the arousal state of an animal at each point in time. This study has important implications for the enormous number of human fMRI studies of functional connectivity, which attempt to quantify the functional relationship between areas based on the correlation of their signals. Specifically, our findings will allow researchers to specifically and quantitatively weigh the effects of arousal on the measured changes in functional connectivity, which has not been done in the past. This method may be of particular importance for investigating patient populations, where arousal changes may be conflated with other changes in the brain. In a two papers, we developed (Papoti et al (2016) Magnetic Resonance in Medicine) and utilized (Hung et al (2015), NeuroImage) radiofrequency coils to measure fMRI signals in the awake marmoset. These are steps in a much larger project that involves a comparison of fMRI signals with local neural activity measured using optical methods. Finally, in a recently published paper, we collaborated with Anil Seth and K. S. Saleem to construct three-dimensional atlas of the monkey brain (Reveley et al (2016), Cerebral Cortex in press). This project will be of great use to the community of nonhuman primate researchers, as it allows for any anatomical scan to be registered and aligned to an atlas depicting the boundaries of cytoarchitectonically defined areas. We are further pursuing two additional projects in this theme. In an ongoing, large project, which is continued from last year, our laboratory is studying the role of the basal forebrain in spontaneous activity. The basal forebrain is a small region that is the origin of many long-range anatomical projections that reach virtually the entire cerebral cortex. Our initial experiments, inactivating portions of this structure in animals undergoing fMRI testing, suggest that the basal forebrain is centrally involved in regulating spontaneous signals throughout the telencephalon. Its effect is particularly pronounced during transitions of arousal, gauged by eye opening and closure. The results from this project are presently in early manuscript form (Turchi, Chang et al, in preparation). In second ongoing project, we are attempting to understand the diversity of neural activity within a fMRI voxel, bridging single-unit data with whole-brain fMRI measurements. In this project, the strategy is to compare the time courses of individual neurons with fMRI voxels throughout the brain during the viewing of movies, which reliably impose activity time courses across the brain. In comparing the time course of each neuron to that of all voxels, it is possible to create functional maps that are specific for individual cells. We reported this method and initial results in last year's Society for Neuroscience meeting (Park et al., SFN Abstr (2015). This project is attempting to link activity the microscopic (single unit) level, directly to the activity in large brain networks, as measured using fMRI. The results from this project are now nearing submission for a manuscript.